I have created a virtual interface (VIF) [Read Linux Bonding / LACP interfaces] using two physical ethernet interfaces of 1GBps each.
When I apply traffic control on the same, I see that the egress bandwidth decreases drastically (at least 10 times lesser). When there is no traffic control, then the egress bandwidth is 10 times higher.
The same traffic control on a physical Ethernet interface does not cause the egress bandwidth to reduce.
Traffic control policy is as below,
$iface is either eth0 or vif; and $limit = 1500000
tc qdisc add dev ${iface} root handle 1: hfsc default 3
tc class add dev ${iface} parent 1:0 classid 1:1 \
hfsc ls rate 10000Mbit ul rate 10000Mbit
tc class add dev ${iface} parent 1:1 classid 1:3 \
hfsc ls rate 10000Mbit ul rate 10000Mbit
tc class add dev ${iface} parent 1:1 classid 1:4 \
hfsc ls rate "${limit}kbit" ul rate "${limit}kbit"
tc filter add dev ${iface} protocol ip parent 1:0 prio 1 \
u32 match ip dport ${port} 0xffff flowid 1:4
Does traffic control on a VIF [LACP interface] cause this issue? I am confused because the above traffic control does not cause any issues on a physical interface. I don't have much idea about the traffic control usage either.
Underlying Base Linux layer is Debian.
